LWT - Food Science and Technology 65 (2016) 645e651

Contents lists available at ScienceDirect

LWT - Food Science and Technology

journal homepage: www.elsevier.com/locate/lwt

Sugar profile, physicochemical and sensory aspects of monofloral

honeys produced by different stingless bee species in Brazilian
semi-arid region
Janaína Maria Batista de Sousa a, Evandro Leite de Souza b, Gilmardes Marques a,

n d, Maria Manuela Pintado d,
Marta de Toledo Benassi c, Beatriz Gullo
a, *
Marciane Magnani
a
Laboratory of Microbial Processes in Foods, Department of Food Engineering, Center of Technology, Federal University of Paraíba, Campus I, 58051-900,
~o Pessoa, Paraíba, Brazil
Joa
b ~o Pessoa,
Laboratory of Food Microbiology, Department of Nutrition, Health Sciences Center, Federal University of Paraíba, Campus I, 58051-900, Joa
Paraíba, Brazil
c
Department of Food Science and Technology, Center of Agrarian Sciences, State University of Londrina, Campus I, 86057-970, Londrina, Parana
, Brazil
d
lica do Porto, Porto, Portugal
Escola Superior de Biotecnologia, Universidade Cato

a r t i c l e i n f o a b s t r a c t

Article history: Monofloral honeys produced by the stingless bee M. Subnida Duke and Melipona scutellaris Latrelle in a
Received 18 June 2015 Brazilian semi-arid region were analyzed regarding their physicochemical (moisture, protein, proline,
Received in revised form hydroxymethyl-furfural, color, electrical conductivity, pH, free acidity, and sugar profile) and sensory
21 August 2015
aspects (color, viscosity and flavor). The floral source influenced the color, acidity, sugar profile, ash and
Accepted 23 August 2015
Available online 28 August 2015
proline content in the honeys. The intensity of the acidic flavor, sweet taste and color in the honeys
varied according to the floral source. No differences were perceived for honeys from the same floral
source produced by different bee species. Principal component analysis revealed that most of the vari-
Keywords:
Floral source
ability was defined by the water content, total acidity, glucose, sucrose and color, as well as by acid taste,
Melipona spp. acid flavor, honey flavor and sweet taste. These findings reveal that Brazilian monofloral honeys pro-
juazeiro duced by stingless bees possess well-defined characteristics that are influenced by the floral source.
malícia © 2015 Elsevier Ltd. All rights reserved.

1. Introduction (Sant’Ana, Sousa, Salgueiro, Lorenzon, & Castro, 2012).

Honey is a natural food consumed without any processing and is
characterized by its complex composition, which varies according
to the bee species, geographical region, available floral source and
storage conditions (Karabagias, Badeka, Kontakos, Karabournioti, &
Kontominas, 2014). The major components of honey are sugars
mostly fructose and glucose although other minor components,
such as enzymes, proteins, organic acids, minerals, pollen grains,
waxes and phytochemicals, are also present (Manzanares, García,
Galdo
n, Rodríguez, & Romero, 2014). Some of these constituents
are naturally found in nectar (or pollen), while others are inserted
by bees during the honey maturation process, thereby defining the
differences in composition and functional properties of the honeys
* Corresponding author.
E-mail address: magnani2@gmail.com.br (M. Magnani).
Traditionally, the combined studies of melissopalinological,
physicochemical and sensory parameters are used to determine the
botanical origin and quality of monofloral honeys. Still, the mineral
content and the sugar profile have been suggested as criteria for the
characterization of monofloral honeys (Manzanares et al., 2014).
The quantitative descriptive analysis (QDA) is usually applied as a
tool to access specific sensory characteristics in honey (Castro-
V
azquez, Díaz-Maroto, Gonz
~ as, & Pe
alez-Vin
rez-Coello, 2009).
However, the descriptive method Ranking Descriptive Analysis
(RDA) has been cited as useful to study sensory properties of foods
because it is less time consuming (in comparison to QDA) and al-
lows, in addition to the descriptions of the attributes, the ranking
among the samples for each assessed sensory attribute (Richter,
Avancini, Prudencio, & Benassi, 2010). Researchers have stated
that application of multivariate statistical methods is a highly
effective tool for characterizing honeys of different geographical,
botanical and entomological origins (Ferreira, Aires, Barreira, &

http://dx.doi.org/10.1016/j.lwt.2015.08.058
0023-6438/© 2015 Elsevier Ltd. All rights reserved.
646 J.M.B. Sousa et al. / LWT - Food Science and Technology 65 (2016) 645e651
Estevinho, 2009; Habib, Al Meqbali, Kamal, Souka, & Ibrahim, 2014;
Silvano, Varela, Palacio, Ruffinengo, & Yamul, 2014).
The semiarid northeastern region of Brazil has a unique biodi-
versity of stingless bees and native flora, enabling the production of
honey with particular characteristics. In this region, a biome (caa-
tinga) composed by botanical species exists that is adapted to the
typical climatic conditions of one rainy and one dry regular season
during the year. Some of the botanical species are commonly visited
by native bee species, known as indigenous bees, stingless bees, or
meliponini bees (Silva et al., 2013, 2014). Among the miliponini
species already identified in the north and northeastern regions of
Brazil, special attention has been given to Melipona subnitida Ducke
(jandaíra) and Melipona scutellaris Latrelle (uruçu); these bee spe-
cies typically visit botanical species available only in a specific
period of the year (rainy or dry season), thus producing different
types of honeys during the year.
Considering these aspects, studies on the characterization of
monofloral honeys produced by native stingless bees in the Bra-
zilian northeastern region could be important incentives for the
maintenance of the stingless bee population and to help conserve
the flora and indigenous bees in this region. In this study, honeys
produced by jandaíra and uruçu bees from specific botanical
sources, which are typically found in the northeastern region of
Brazil during the rainy and dry seasons, were characterized in terms
of their sugar content, physicochemical parameters and sensory
attributes.
2. Material and methods
2.1. Experimental design and honey samples
The experimental design included four different monofloral
honeys produced by two different stingless bee species in the
semiarid region of Brazilian northeastern (Serido
region, state of
Rio Grande do Norte and Agreste region of state of Paraíba)
collected in three different occasions (4
2
3). Each of the 24
samples analyzed was composed by a mixture of honeys collected
in four different meliponaries. The samples of honey from Ziziphus
joazeiro Mart. (juazeiro) and Mimosa quadrivalvis L. (malícia) pro-
duced by the stingless bees M. subnitida Ducke (jandaíra) and
M. scutellaris Latrelle (uruçu) were collected during the 2012 dry
season, while the honey samples from Mimosa arenosa willd Poir
(jurema branca) and Croton heliotropiifolius Kunth (velame branco)
produced by the both bee species (jandaíra and uruçu) were
collected during the 2013 rainy season. Samples were stored in
sterilized amber glass containers, shipped to the laboratory, and
then maintained at 6e8  C in the dark until analysis. To ensure that
the honeys were monofloral, the melissopalynological procedure
was applied to all samples (Silva et al., 2013). For this procedure,
10 g of each honey sample was diluted in 20 mL of distilled water
and then centrifuged at 4000 g for 20 min. The sediment was dried
at 40  C and then mounted with Entellan Rapid (Merck,
1.07961.0500). The honeydew elements and pollen grains (n ¼ 500)
were counted and identified in 20 distinct optical areas using an
optical microscope (Nikon Optiphot II microsco
pio; 400
and
1000
). The pollen grains were compared to reference images of
University of Sa~o Paulo, S~ao Paulo, Brazil. All samples contained
more than 80% pollen grains of the same botanical origin.
2.2. Physicochemical analysis
All of the physicochemical parameters were determined in
triplicate according to the Harmonized Methods of the
International Honey Commission e IHC (2002) and the Association
of Official Analytical Chemists e AOAC (1990). The water content
was determined via refractometry using an Abbe
-type refractom-
eter at 20  C (Q767-B, Tokyo, Japan); the corresponding values were
obtained according to the table of Chatway. The brix values in
honeys were determined via refractometry following the AOAC
Official method 31.119. The total acidity was analyzed using the sum
of free acidity and lactonic acidity determined by titration of 10 g of
honey dissolved in 75 mL of distilled water with 0.05 mol/L NaOH
to pH 8.3, with the results expressed as mmol Hþ/L. The pH was
measured in a solution of 10 g of honey in 75 mL ultrapure water
using a digital potentiometer (QUIMIS, Q488AS). The protein con-
tent was measured using the Kjeldahl method, based on the con-
version of the organic nitrogen present in the sample to (NH4)2SO4,
using 6.25 as the conversion factor. The proline content was
determined via the measurement of the absorbance at 510 nm of
the resulting product between proline and ninhydrin in an acidic
medium. The electrical conductivity was measured at 20  C in a
20 g/100 mL (w/v) solution of honey in ultrapure water using a
radiometer (Analyser, model 600), with the results expressed as
mS/cm. The ash content was measured after burning the sample at
550  C. The color was determined using the Pfund method; the
absorbance of the honeys [diluted to 50 g/100 mL (w/v) in ultrapure
water] was determined at 636 nm and then converted to the Pfund
scale (mmPfund ¼ 38.7 þ 371.39
Abs). Hydroxymethyl furfural
(HMF) was measured using the method of White (Silvano et al.
2014).
2.3. Sugar profile
Approximately 0.5 g of honey was weighed directly in poly-
propylene tubes and then mixed with 10 mL of water. Next, a
milliliter of this dissolution was passed through a 0.20-mm cellulose
acetate membrane prior to HPLC analysis. The determination of the
sugars (fructose, glucose, sucrose, maltose and arabinose) was
performed using a 1100 series HewlettePackard chromatograph
equipped with a refractive index detector operated at 50  C and a
300
7.8 mm CARBOsep CHO 682 column (Transgenomic, Glas-
gow, U.K.) operating at 80  C. Distilled water was used as the mobile
phase (flow rate 0.4 mL/min). The HPLC sample peaks were iden-
tified by comparing the retention times obtained from standards.
The honey samples were also spiked with standards (Sigma®) to
verify the identity of the chromatographic peaks. Duplicate in-
jections were performed, and the average peak areas were used for
the quantification.
2.4. Sensory analysis
This work was submitted to the National Committee in
Research Ethics (Brazil) and approved under number CAAE:
06371012.8.0000.5188. Sensory analyses were performed using
Ranking Descriptive Analysis (RDA) according to the procedures
described by Richter et al. (2010). The panel was composed of 19
assessors, all members (students, technicians and teachers) of the
Federal University of Paraíba (Joa ~o Pessoa, Brazil), with ages
ranging from 16 to 62 years and with previous experience in sen-
sory analysis. The sessions were performed in a tasting room,
separated from the area where the samples were prepared and
maintained at room temperature (25  C) during the evaluation. The
coded samples were presented in random order to each assessor in
40-mL glass vials sealed with a twist-off cap and at room temper-
ature. The order of presentation was balanced and randomized to
eliminate contrast effect and positional bias. Mineral water was
provided for the assessors to perform mouth-rinsing between
samples. The sessions were performed between 10:00 and 12:00,
and the panelists declared that they were not hungry at the
commencement of the tests. During the initial sessions, the
 J.M.B. Sousa et al. / LWT - Food Science and Technology 65 (2016) 645e651 647

assessors underwent training in descriptive analysis, during which
they generated descriptors over the course of seven sessions. Due to
the high number of honey samples, the panel opted to evaluate four
samples per session (honeys of juazeiro and malícia produced by
jandaíra and uruçu bee) and (honeys of jurema branca and velame
branco produced by jandaíra and uruçu bee). The common de-
scriptors chosen by the panel were compiled, together with sig-
nificant descriptors selected for the formulation of a glossary
(amber color, viscosity, honey flavor, acid flavor sweet taste, bitter
taste, and acid taste). The honey samples were evaluated by RDA in
two distinct sessions, where each sample was individually evalu-
ated using the glossary and then ordered according to increasing
intensity for each of the attributes.
honeys from the same floral source produced by different bee
species (Table 1), with the exception of honey from jurema branca.
In honey, water content has been linked to the floral source,
although climatic conditions, soil, collection period and processing
aspects may strongly affect this parameter (Silvano et al., 2014). The
water content values found in the assessed honeys are similar to
those reported in previous studies (22e35 g/100 g) involving
characterization of honeys produced by jandaíra and uruçu bees in
different states of northeastern Brazil (Silva et al., 2013; Sousa et al.,
2013; Souza et al., 2006), which indicate that the analyzed samples
presented satisfactory quality.

3.2. Brix

2.5. Data analysis
All analyses were performed in triplicate in three different
occasions. The mean values obtained for the physicochemical
variables studied in honey samples were compared using the
The  Brix values for the honey samples ranged from 71.1 to 74.7.
No differences (p  0.05) were observed regarding the floral source
among the assessed honey samples (Table 1). However, honeys
from malícia differed (p  0.05) according to the bee species. The
 Brix values are directly related to the levels of sugars present in

Two-Way ANOVA followed by Tukey test. The sensory data were
analyzed by the Friedman test, considering four samples, 19
judges and the critical value: 21. Principal Component Analysis
(PCA) was performed to integrate significant data. For this, the
physicochemical and sensory data that differed among honeys
from distinct floral sources, as well as those considered charac-
teristic of honeys produced by stingless bee, such as water con-
honey, making them an important indicator of possible adultera-
tion (Guarrini et al., 2009). Generally, honeys produced by stingless
bees exhibit lower  Brix values when compared to honeys pro-
duced by A. mellifera (75), as honeys from stingless bees possess
higher water content and a lower percentage of total sugars
(Guarrini et al., 2009; Habib et al., 2014). Previous studies found
 Brix values similar to those detected in this study for honeys

tent and acidity, were normalized and submitted to PCA produced by the uruçu bee in the state of Paraíba - Brazil (average of
multivariate analysis. All statistical analyses were performed us- 72  Brix) (Campos, Gois, & Carneiro, 2010), and for honeys pro-
ing Statistica Software 5.1. A p-value  0.05 was considered to be duced by stingless bees in South American countries (57.5e75.0
significant.  Brix) (Souza et al., 2006).

3. Results and discussion 3.3. pH and acidity

3.1. Water content (g/100 g) The pH values found in honeys ranged from 3.1 to 5.3, differing
(p  0.05) among the honeys from different floral sources and/or
The water content of the honey samples ranged from 23.9 to bee species. Previous studies describe pH values less than and
28.9 g/100 g, and no differences were observed (p  0.05) among greater than (2.9e6.5) those found in this study for multifloral

Table 1
Physicochemical parameters (n:3, mean values ± standard deviation) and the sugar profiles of monofloral honeys produced by different stingless bees (jandaíra and uruçu) in
the semiarid region of Northeastern Brazil from different floral sources.

Physicochemical parameters Stingless bee/monofloral honeys

jandaíra (M. Subnida Duke) uruçu (M. scutellaris Latrelle)

juazeiro malícia velame branco jurema branca juazeiro malícia velame branco jurema branca

Water (g/100 g) 23.9Ad ± 0.4 27.2Ab ± 0.2 25.6Ac ± 0.4 28.9Aa ± 0.2 24.3Ab ± 0.3 26.5Aa ± 0.8 25.8Aa ± 0.4 25.4Bab ± 0.6

Brix 74.7Aa ± 0.2 71.1Bb ± 0.2 72.4Aab ± 1.2 72.0Aab ± 1.0 74.3Aa ± 0.5 73.0Aa ± 0.6 72.5Aa ± 0.5 72.8Aa ± 1.2
pH 5.3Aa ± 0.4 3.1Bc ± 0.2 3.8Ab ± 0.01 3.6Ab ± 0.02 4.2Ba ± 0.05 4.0Ab ± 0.02 3.5Bc ± 0.4 3.6Ac ± 0.1
TA (mmol Hþ/L) 28.2Ac ± 0.9 86.8Aa ± 0.4 17.8Bd ± 1.0 37.8Bb ± 0.8 30.4Ac ± 1.2 66.1Ba ± 2.4 32.1Ac ± 0.4 42.4Ab ± 0.7
Protein (g/100 g) 0.5Aa ± 0.1 0.4Aab ± 0.1 0.3Aa ± 0.0 0.2Aa ± 0.1 0.5Aa ± 0.0 0.4Aab ± 0.0 0.3Ab ± 0.0 0.3Ab ± 0.0
Proline (mg/kg) 20.5Aa ± 3.7 11.9Ab ± 2.1 5.9Ab ± 0.8 10.8Ab ± 1.1 17.4Aa ± 0.9 7.5Bb ± 0.5 4.6Ac ± 1.0 8.9Ab ± 0.7
EC (mS/cm) 598Bb ± 2.6 636Aa ± 1.0 300Bd ± 3.1 520Bc ± 0.2 670Aa ± 5.1 514Bc ± 2.3 340Ad ± 1.5 571Ab ± 3.4
Ash (g/100 g) 0.52Aa ± 0.0 0.04Ab ± 0.0 0.11Ab ± 0.0 0.12Ab ± 0.0 0.41Aa ± 0.0 0.03Bb ± 0.0 0.12Ab ± 0.0 0.1Ab ± 0.0
Color (nm Pfund) 95.4Ba ± 0.2 55.6Bb ± 0.02 35.8Bd ± 0.3 54.1Bc ± 0.1 103.4Aa ± 0.6 82.8Ab ± 0.4 55.9Ac ± 0.4 57.2Ac ± 1.1
HMF nd nd nd nd nd nd nd nd
Sugars (g/100 g)
Glucose (g/100 g) 37.7Ac ± 0.4 45.4Aa ± 0.3 42.1Bb ± 0.6 45.7Aa ± 0.4 38.1Ac ± 0.3 42.6Ba ± 0.2 43.3Aa ± 0.1 41.4Bb ± 0.5
Fructose (g/100 g) 59.2Aa ± 1.1 50.0Bd ± 0.9 55.7Ab ± 1.9 52.6Ac ± 2.5 57.6Aa ± 1.5 55.5Aab ± 1.0 53.8Bb ± 0.9 53.6Ab ± 0.7
Sucrose (g/100 g) 1.6Ab ± 0.2 3.9Aa ± 0.2 0.7Bc ± 0.0 1.2Bbc ± 0.2 2.6Ba ± 0.5 1.9Ba ± 0.3 2.0Aa ± 0.3 3.0Aa ± 0.5
Maltose (g/100 g) nd nd nd nd nd nd nd nd
Arabinose (g/100 g) nd 0.5A ± 0.0 1.0A ± 0.1 0.4A ± 0.1 0.3b ± 0.0 0.4Ab ± 0.1 0.7Ba ± 0.1 0.2Ac ± 0.0
TS 68.2Aab ± 0.7 63.2c ± 1.0 67.9Abc ± 1.9 71.2Aa ± 1.4 67.6Aab ± 1.6 62.7Ab ± 1.4 68.0Aab ± 2.3 71.2Aa ± 1.5
F/G 1.5Aa ± 0.2 1.1Bc ± 0.0 1.3Ac ± 0.0 1.1Bc ± 0.0 1.5Aa ± 0.0 1.3Ab ± 0.0 1.2Bb ± 0.0 1.2Ab ± 0.0

TA ¼ Total acidity; EC ¼ Electrical conductivity; TS ¼ Total Sugars; nd ¼ not detected.: (Ziziphus joazeiro); malícia: (Mimosa quadrivalvis); Velame branco: (Croton helio-
tropiifolius); Jurema branca: (Mimosa arenosa). Different capital letters in the same row indicates significant differences (p  0.05) between honeys from the same botanical
source produced by distinct bee species according to the Tukey test. Different lowercase letters in the same row indicates significant differences (p  0.05) between honeys
from different floral source produced by the same bee species according to the Tukey test.
648 J.M.B. Sousa et al. / LWT - Food Science and Technology 65 (2016) 645e651

honeys produced by stingless bees (Sousa et al., 2013). In general,
honeys from a floral source or honeydew honeys exhibit acidic pH
values, without a direct relationship with the botanical source.
However, the pH values in honeys have been linked with other
parameters that are influenced by geographical and botanical
origin, such as mineral content (Vanhanen, Emmertz & Savage,
2011).
The acidity values ranged from 17.88 to 86.85 mmol Hþ/L in
the samples evaluated. Differences were observed (p  0.05)
among the honeys from different floral sources and/or bee spe-
cies (Table 1). Previous studies that characterized wild honey
produced by stingless bees in north and northeastern Brazil re-
ported lower and higher acidity values (18.5e95.2 mmol Hþ/L)
when compared to those observed in the present study (Alves,
Carvalho, Souza, Sodre
, & Marchini, 2005; Sousa et al., 2013).
The acidity value corresponds to the balance of organic acids
present in honey, (Vit, 2008), which vary according to the floral
composition and the bee species. This variation may explain the
differences observed.
3.4. Protein and proline
The mean protein values in honeys ranged from 0.2 to 0.5 g/
100 g (Table 1), and no difference (p > 0.05) in this parameter was
observed among the honeys from the same floral source, despite
being produced by bees from different species. These results rein-
force the findings of Habib et al. (2014), who revealed that the
protein content in honeys is dependent on the floral source and
that different values may be associated with the presence of pro-
teins derived from nectar of flowers and enzymes introduced by
bees in honeys.
The proline levels in honey ranged from 20.5 to 4.6 mg/kg
(Table 2). The honeys from juazeiro produced by both bee species
also exhibited higher proline contents (p  0.05) when compared
to the other honeys analyzed; however, honeys from the same
floral source exhibited similar proline contents (p  0.05). Proline is
the main free amino acid present in honeys produced by stingless
bees, making it useful for the characterization of the botanical
source because is related to the floral source and the amount of
pollen present in the honeys (Truzzi, Annibaldi, Illuminati, Finale, &
Scarponi, 2014).
3.5. Sugar quantitative profile
The type of sugar present in the greatest amounts in the honeys
was fructose, followed by glucose and sucrose (Table 1). With the
exception of honeys from juazeiro, the honeys of the same floral
source produced by different bee species differed (p  0.05) in the
amounts of each detected sugar; this difference was also found for
different floral sources. In the floral honeys, the glucose content
corresponds to the characteristics of the nectar of the predominant
flowers and varies among the plant species (Escuredo, Dobre,
Fern
andez-Gonz
alez, & Seijo, 2014). The glucose and fructose
levels of the honeys studied were higher than those reported in
previous studies, with multifloral honeys produced by stingless
bees in Ecuador (25.5 g/100 g of total sugars for glucose and 25.2 g/
100 g of total sugars for fructose) (Guerrini et al., 2009) and in
southern Brazil (8.2e35.39 g/100 g of total sugars for glucose and
31.88e45.46 g/100 g of total sugars for fructose) (Rize
lio et al.,
2011).
The fructose/glucose (F/G) ratio in the honeys evaluated ranged
from 1.1 to 1.5 (Table 1), which is similar to the ratio found by Oddo
et al. (2008) (1.4) in honey samples produced by stingless bees in
Australia. The F/G ratio directly influences the sweet taste of honey
because fructose is sweeter than glucose. The sucrose contents
found are lower than those described in previous studies for mul-
tifloral honeys produced by the jandaíra bee (3.7 g/100 g of total
sugars) and the uruçu bee (5.3e8.8 g/100 g of total sugars) (Campos
et al., 2010; Souza et al., 2013). These results demonstrate the lack
of adulteration and honey collection at the ideal maturation time
because high sucrose content may result from the addition of
commercial sugar or may be attributed to the early honey
collection.
3.6. Electrical conductivity
The electrical conductivity (EC) of the honeys studied ranged
from 300 to 670 mS/cm(Table 1), classifying them as floral honeys
(800 mS/cm) (Codex Alimentarius, 2001). Differences were found
in the EC (p  0.05), regardless of the floral source or bee species.

Table 2
Intensity values of the sensory attributes (assessed by RDA) of monofloral honeys produced by different stingless bees (jandaíra and uruçu) in the semiarid region of
Northeastern Brazil from different floral sources.

Sensory attributes Stingless bee/monofloral honeys

Jandaíra (M. subnida Duke) Uruçu (M. scutellaris Latrelle)

juazeiro malícia juazeiro malícia

Amber color 62a 22b 71a 38b

Viscosity 71a 20b 61a 38b
Honey flavor 68a 29b 58a 35b
Acid flavor 44b 72a 21c 53ab
Sweet taste 62a 16c 69a 43b
Acid taste 40b 74a 19c 54ab
Bitter taste 52a 60a 56a 20b

velame branco jurema branca velame branco jurema branca

c ab cb
Amber color 26 56 37 71a
Viscosity 55a 44ab 64a 27b
Honey flavor 29c 56ab 39 bc 66a
Acid flavor 58ab 28c 65a 39bc
Sweet taste 49a 66a 25b 50a
Acid taste 52ab 43bc 71a 21c
Bitter taste 59a 36b 44 ab 51ab

Rank sum values of answers of 19 assessors for the ordination of intensity of assessed sensory attributes in honey samples, considering four samples, 19 judges and the critical
value: 21. Different letters in the same row indicate significant differences (p  0.05) according to Newell and MacFarlane (1987); juazeiro: (Ziziphus joazeiro); malicia:
(Mimosa quadrivalvis); velame branco: (Croton heliotropiifolius); jurema branca: (Mimosa arenosa).
 J.M.B. Sousa et al. / LWT - Food Science and Technology 65 (2016) 645e651
Campos et al. (2010) observed values from 539.6 to 586 mS/cm for
multifloral honeys produced by uruçu bees in a semiarid region of
northeastern Brazil. Results similar to those of the present study
were previously reported for monofloral honeys produced by
A. mellifera in a semiarid region of the United Arab Emirates
(254e605 mS cm1) (Habib et al. 2014). The electrical conductivity
intensity of honeys has been described as a measure of the amount
of mineral elements possessing good electrical conductivity prop-
erty contained in the honey (Yadata, 2014). However, higher elec-
trical conductivity values do not necessarily correspond to higher
amounts of ash in honeys (Escuredo et al., 2014), as evidenced in
the present study. Other factors, such as floral source, the amount of
organic acids and proteins, and storage time can also influence the
electric conductivity of honey samples (Karabagias et al., 2014).
3.7. Ash
Honeys of the same botanical origin did not differ (p  0.05) for
different producer bee species (Table 1). The highest ash values
(p  0.05) were observed for honeys from juazeiro. Ash content,
representing the total minerals present in honey, is directly related
to the environmental, geographical and botanical aspects of the
region where it is produced (Finola, Lasagno, & Marioli, 2007). The
ash content observed for honeys from malícia, velame branco and
jurema branca are similar to the levels reported for honeys from
angico (Anadenanthera sp.) (0.01e0.02 g/100 g) produced by jan-
daíra bees (Silva et al., 2014) and for wild honeys (0.1e0.302 g/
100 g) produced by uruçu bees (Campos et al., 2010) in a Brazilian
semiarid region, reinforcing the influence of geographic factors on
this parameter (Santos, Moreira, & Maria, 2015). The high ash
content in some honey may be due to the nectar characteristics of
some botanical species, such as juazeiro in the present study.
3.8. Color and HMF
The color of honeys on the Pfund scale exhibited values from
35.8 to 103.4 (Table 2). Honeys from juazeiro produced by both bee
species exhibited the highest values (p < 0.05) on the Pfund scale,
being classified as amber. Interestingly, these honeys also exhibited
the highest ash and pH values (p < 0.05) when compared to all
other honey samples. Previous studies have reported that the color
intensity of honeys varies according to pH and mineral content,
exhibiting a positive correlation with these two parameters
(Guerrini et al., 2009). However, factors such as exposure to light,
heat and storage time, as well as enzymatic reactions, may also
affect this parameter.
HMF was not detected in honey samples characterizing the
evaluated monofloral honeys from stingless bees as fresh and high
quality honeys (Habib et al., 2014). HMF is widely known as a
parameter of freshness in honeys because it is absent in fresh
honeys and tends to increase during processing and/or aging of the
product.
3.9. Sensory analysis
The sensory characteristics of aroma and acidic flavor, sweet
taste and intense amber color (caramel) described for monofloral
honeys are reported in the sensory evaluation of honeys produced
by stingless bees (Ferreira et al., 2009; Souza et al., 2013); however,
there are no previous studies on the ordering of these attributes
according to the perceived intensity. In general, honeys from the
same floral source were observed to have similar sensory attributes
and did not differ with respect to the bee species that produced
them. The physicochemical and sensory characteristics of honeys
vary according to botanical origin, maturation time, weather and
649
storage conditions (Ferreira et al., 2009).
Considering the ordering of the intensity of the attributes of
amber color, viscosity, honey aroma and sweet taste, the judges
determined that the honeys from juazeiro produced by both spe-
cies of stingless bees were more intense (p  0.05) compared with
the honeys from malícia (Table 2). Previous studies have reported a
direct correlation of these sensory attributes in honeys with
physicochemical parameters, such as  Brix, pH, ash content, pro-
teins, color and total sugars (Habib et al., 2014). This correlation
may explain the difference in intensity of the attributes of honeys
from juazeiro and malícia perceived by the judges in the evaluation
because the  Brix values, protein, ash, total sugar and instrumental
color were higher (p  0.05) in the honeys from juazeiro compared
to those in the honeys from malícia for both bee species (Table 2).
This result is also corroborated by the fact that the judges deter-
mined the honeys from malícia as being more acidic (aroma and
acid taste) than those from juazeiro (p  0.05); in addition, the
judges determined that the honeys from malícia did not differ
(p  0.05) in aroma and acid taste when considering the bee species
(Table 2). The greater intensity of aroma and acid taste described in
the sensory evaluation for honeys from malícia compared to those
from juazeiro appears to be associated with their higher total
acidity, lower pH index, and the lowest total sugar, which could
mask the acidity (Table 1). These features are possibly derived from
the nectar offered by the malícia flower to the bee species,
considering that acidity is a parameter strongly related to the
botanical origin (Beay, Solomon, Bultossa, Adgaba, & Melaku, 2015).
Similar to the observed intensity ordering of the attributes of
honeys from juazeiro and malícia, honeys from jurema branca were
perceived as being more intense (p  0.05) regarding the attributes
of color, viscosity, honey aroma and sweet taste when compared to
honey from velame branco (Table 2), but were not considered to be
different (p > 0.05) when considering the bee species. In the
physicochemical analysis, honeys from velame branco exhibited
high values for  Brix, pH, ash, protein, instrumental color and total
sugars (Table 1), partially explaining the sensory characterization.
In contrast, higher intensities for the attributes of acid aroma, acid
taste and bitter taste were described for honeys from velame
branco (Table 2), when compared to honeys from jurema branca.
Although these attributes are directly correlated with the total
acidity levels, the ash content, and the pH and type of organic acids
present in monofloral honeys (Manzanares et al., 2014), high acidity
is typical physicochemical and sensory characteristic of honeys
produced by stingless bees (Vit, 2008). In the present study, honeys
from malícia and velame branco exhibited higher intensity for the
attributes of aroma and acid taste and lower intensity for amber
color when compared to honeys from juazeiro and jurema branca.
This result is in agreement with the previous findings that honeys
produced by stingless bees that exhibit less intense amber color
have more intense aroma and acid taste (Ferreira et al., 2009).
Once the individual behavior of each honey was studied, a PCA
was used to assess the overall effect of the variables (floral source
and bee specie) on the principal components that define the honey
samples. The physicochemical parameters of the honey samples
that provide a major contribution to PC1 were water content, total
acidity, glucose and color; sensory parameters that contribute in
this PC were acid taste and acid flavor together with honey flavor
and sweet taste, which were in contrast to the aforementioned
result (Fig. 1A). PC1, which explains 48.15% of the variance, clearly
separates acid taste and acid flavor from sweet taste, water content,
glucose and total acidity. However, PC2, which explains 17.14% of
the variance, was only defined by sucrose level; the remaining
variability not taken into account by PC1 was explained by this
variable. For PC1, the water content, total acidity and glucose were
found to be the most important variables that explain the
650 J.M.B. Sousa et al. / LWT - Food Science and Technology 65 (2016) 645e651

Fig. 1. (A) Principal Component Analysis (PCA) of the physicochemical (water content, total acidity, glucose, sucrose) and sensory (acid flavor, acid taste, viscosity, sweet taste, bitter
taste, honey flavor, color) aspects of monofloral honey; (B) Distribution of samples according PCA: JJ (juazeiro/jandaíra); UJ (juazeiro/uruçu); JA (malícia/jandaíra); UA (malícia/
uruçu); JJB (jurema branca/jandaíra); UJB (jurema branca/uruçu); JVB (velame branco/jandaíra); UVB (velame branco/uruçu).

separation of the honey of malícia produced by both bee species
studied from the other tested honeys (Fig. 1B).
Previous studies also described the characteristics of acid flavor
and acid taste for the honey of malícia produced in a semiarid re-
gion of Brazil (Campos et al., 2010; Sousa et al., 2013), which are
strongly associated with the nectar of the floral source collected by
bees (Belay et al., 2015; Silvano et al., 2014). PC1 also explains the
separation of the honeys of juazeiro produced by uruçu bees from
the other honeys assessed (Fig. 1B), which could be related to the
higher values of color, honey flavor and sweet taste when
compared with the honeys of the other floral sources studied
(Table 2). The honey of juazeiro produced by jandaíra bees also
exhibited higher values for these same parameters (color, honey
flavor and sweet taste); note that this honey was ranked by the
 J.M.B. Sousa et al. / LWT - Food Science and Technology 65 (2016) 645e651
sensory panel as being more intense regarding acid taste and acid
flavor, which could influence the distribution of the samples ac-
cording to the PCA (Table 2; Fig. 1B). However, the honey of juazeiro
exhibited no differences (p  0.05) when produced by different bee
species regarding the physicochemical parameter total acidity.
According to PC2, the similarity observed between the honey of
jurema branca and velame branco produced by jandaíra was
defined by the levels of sucrose (Fig. 1B). These samples exhibited
the lowest sucrose amounts when compared to the other assessed
honey samples (Table 1). This result suggests that this variable
strongly contributes, with the same weight, despite the floral
source, for the unique characteristics in these two honeys, which
appears in the y axis (Fig. 1B); this result is similar to results found
for the honeys of jurema branca and velame branco produced by
jandaíra in the sensory evaluation by RDA (Table 2). For the honey
samples studied, the water content exhibited a positive correlation
with glucose (r ¼ 0.88) and with total acidity (r ¼ 0.5), whereas a
negative correlation was observed with viscosity (r ¼ 0.54) and
color (r ¼ 0.50). As expected, honeys with lower water content
(juazeiro and velame branco) were ranked as being more intense
for the viscosity attribute when compared to the other honeys
(malícia and jurema branca), regardless of the bee species involved.
A strong positive correlation (r ¼ 0.89) was also observed for acid
flavor and acid taste, revealing the accuracy of the panel regarding
the sensory evaluation performed. In contrast, a negative correla-
tion was observed between total acidity and sweet taste
(r ¼ 0.63). The highest total acidity values were observed for
malícia honey produced by both bee species assessed; this honey
exhibited lower values for the intensity of sweet taste in the sen-
sory evaluation by RDA.
4. Conclusions
Monofloral honeys produced by distinct stingless bee species in
a Brazilian northeastern region presented unique characteristics
related to the floral source and a good correlation between the
physicochemical and sensory parameters. These findings revealed
that RDA is a feasible technique for establishing differences be-
tween honey samples. In particular, honeys from juazeiro differed
from other honeys regarding their intense color and high proline,
glucose, and ash contents, while honeys from malícia exhibited the
highest acidity.
Acknowledgments
Thank CAPES- and CNPq-Brazil and the beekeepers for
providing the honey samples.
References

, G. S., & Marchini, L. C. (2005).
Alves, R. M. O., Carvalho, C. A. L., Souza, B. A., Sodre
Physico-chemical characteristics of honey samples of stingless bee Melipona
mandaçaia Smith (Hymenoptera:Apidae). Ci^ encia e Tecnologia de Alimentos, 25,
644e650.
AOAC (Association of Official Analytical Chemists). (1990). Official methods of the
analysis of official analytical chemists (15th ed., vol. II) USA: Virginia.
Belay, A., Solomon, W. K., Bultossa, G., Adgaba, N., & Melaku, S. (2015). Botanical
origin, colour, granulation, and sensory properties of the Harenna forest honey,
Bale, Ethiopia. Food Chemistry, 167, 213e219.
Campos, F. S., Gois, G. C., & Carneiro, G. G. (2010). Physico-chemical parameters of
651
the honey of stingless bee Melipona scutellaris produced in the Paraíba. Zoo-
tecnia, 7, 186e190.

zquez, L., Díaz-Maroto, M. C., Gonza
lez-Vin
~ as, M. A., & Pe

rez-Coello, M. S.
Castro-Va
(2009). Differentiation of monofloral citrus, rosemary, eucalyptus, lavender,
thyme and heather honeys based on volatile composition and sensory
descriptive analysis. Food Chemistry, 12, 1022e1030.
Codex Alimentarius. (2001). Commission Standards. Official methods of analysis.
CODEX STAN. 12e1981, Rev. 1 1987, Rev. 2.
Escuredo, O., Dobre, I., Fern
andez-Gonza
lez, M., & Seijo, M. C. (2014). Contribution
of botanical origin and sugar composition of honeys on the crystallization
phenomenon. Food Chemistry, 149, 84e90.
Ferreira, I. C. F. R., Aires, E., Barreira, J. C. M., & Estevinho, L. M. (2009). Antioxidant
activity of Portuguese honey samples: different contributions of the entire
honey and phenolic extract. Food Chemistry, 114, 1438e1443.
Finola, M. S., Lasagno, M. C., & Marioli, J. M. (2007). Microbiological and chemical
characterization of honeys from central Argentina. Food Chemistry, 100,
1649e1653.
Guerrini, A., Bruni, R., Maietti, S., Poli, F., Rossi, D., Paganetto, G., et al. (2009).
Ecuadorian stingless bee (Meliponinae) honey: a chemical and functional
profile of an ancient health product. Food Chemistry, 114, 1413e1420.
Habib, H. M., Al Meqbali, F. T., Kamal, H., Souka, U. D., & Ibrahim, W. H. (2014).
Physicochemical and biochemical properties of honeys from arid regions. Food
Chemistry, 153, 35e43.
International Honey Commission e IHC. (2002). Swiss Bee Research Centre. FAM,
Liebefeld, CH-3003 Bern, Switzerland.
Karabagias, I. K., Badeka, A. V., Kontakos, S., Karabournioti, S., & Kontominas, M. G.
(2014). Botanical discrimination of Greek unifloral honeys with physico-
chemical and chemometric analyses. Food Chemistry, 165, 181e190.
Manzanares, A. B., García, Z. H. B., Galdo
n, R., Rodríguez, E. R., & Romero, C. D.
(2014). Physicochemical characteristics of minor monofloral honeys from Ten-
erife, Spain. LWT Food Science and Technology, 55, 572e578.
Newell, G. J., & Macfarlane, J. D. (1987). Expanded tables for multiple comparison
procedures in the analysis of ranked data. Journal of Food Science, 52,
1721e1725.
Oddo, P. L., Heard, T. A., Rodríguez-Malaver, A., Pe
rez, R. A., Ferna

ndez-Muin~ o, M. A.,
Sancho, M. T., et al. (2008). Composition and antioxidant activity of trigona
carbonaria honey from Australia. Journal of Medicinal Food, 11, 789e794.
Richter, V. B., Avancini, T. C., Prudencio, S. H., & Benassi, M. T. (2010). Proposing a
ranking descriptive sensory method. Food Quality and Preference, 21, 611e620.
Rize
lio, V. M., Tenfen, L., Silveira, R., Gonzaga, L. V., Costa, A. C. O., & Fett, R. (2011).
Development of a fast capillary electrophoresis method for determination of
carbohydrates in honey samples. Talanta, 93, 62e66.
Santos, A., Moreira, R. F. A., & Maria, C. A. B. (2015). Study of the principal con-
stituents of tropical angico (Anadenanthera sp.) honey from the atlantic forest.
Food Chemistry, 171, 421e425.
Sant’Ana, L. D., Sousa, J. P. L. M., Salgueiro, F. B., Lorenzon, M. C. A., & Castro, R. M.
(2012). Characterization of monofloral honeys with multivariate analysis of
their chemical profile and antioxidant activity. Journal of Food Science, 77,
135e140.
Silvano, M. F., Varela, M. S., Palacio, M. A., Ruffinengo, S., & Yamul, D. K. (2014).
Physicochemical parameters and sensory properties of honeys from Buenos
Aires region. Food Chemistry, 152, 500e507.
Silva, T. M. S., Santos, F. P., Evangelista-Rodrigues, A., Silva, E. M. S., Silva, G. S.,
Novais, J. S., et al. (2013). Phenolic compounds, melissopalynological, physico-
chemical analysis and antioxidant activity of jandaíra (Melipona subnitida)
honey. Journal of Food Composition and Analysis, 29, 10e18.
Silva, T. M. G., Silva, P. R., Camara, C. A., Silva, G. S., Santos, F. A. R., & Silva, T. M. S.
(2014). Chemical analysis and antioxidant potential of angico honey collected
by stingless bee Jandaíra. Journal of Chemistry, 5, 1370e1379.
Sousa, J. M. B., Aquino, I. S., Magnani, M., Albuquerque, J. R., Santos, G. G., &
Souza, L. S. (2013). Physicochemical aspects and sensory profile of stingless bee
honeys from Serido
region, State of Rio Grande do Norte, Brazil. Semina: Ci^ encias
Agra
rias, 34, 1763e1772.
Souza, B., Roubik, D., Barth, O., Heard, T., Enríquez, E., Carvalho, C., et al. (2006).
Composition of stingless bee honey: setting quality standards. Interci^ encia, 31,
867e875.
Truzzi, C., Annibaldi, A., Illuminati, S., Finale, C., & Scarponi, G. (2014). Determina-
tion of proline in honey: comparison between official methods, optimization
and validation of the analytical methodology. Food Chemistry, 150, 477e481.
Vanhanen, L. P., Emmertz, A., & Savage, G. P. (2011). Mineral analysis of mono-floral
New Zealand honey. Food Chemistry, 128, 236e240.
Vit, P. (2008). Review: valorization honey of stingless bees (Meliponini). Brazilian
Journal of Pharmaceutical Science, 50, 20e28.
Yadata, D. (2014). Detection of the electrical conductivity and acidity of honey from
different areas of Tepi. Food Science and Technology, 2, 59e63.